Você está na página 1de 5

1

Some issues with Quasi-Steady State Model in


Long-term Stability

arXiv:1310.0058v1 [cs.SY] 30 Sep 2013

Xiaozhe Wang, Hsiao-Dong Chiang, Fellow.

AbstractThe Quasi Steady-State (QSS) model of long-term


dynamics relies on the idea of time-scale decomposition. Assuming that the fast variables are infinitely fast and are stable in
the long-term, the QSS model replaces the differential equations
of transient dynamics by their equilibrium equations to reduce
complexity and increase computation efficiency. Although the
idea of QSS model is intuitive, its theoretical foundation has
not yet been developed. In this paper, several counter examples
in which the QSS model fails to provide a correct approximation
of the complete dynamic model in power system are presented
and the reasons of the failure are explained from the viewpoint
of nonlinear analysis.
Index Termsquasi-steady state model, complete dynamical
model, long-term stability.

I. I NTRODUCTION

He ever-increasing loading of transmission networks together with a steady increase in load demands has pushed
the operation conditions of many power systems ever closer
to their stability limits [1]- [5]. Voltage stability has become
one of the major concerns for the secure operation of power
systems. Voltage stabilities are classified into transient voltage
stability, mid-term stability and long-term stability based on
different time scales. The distinction between mid-term and
long-term can be based on neither fixed time-frame basis nor
modelling requirements [2], hence we only use long-term time
scale in this paper to denote the one beyond the transient time
scale for stability analysis. This paper considers the long-term
voltage stability model.
Power system dynamic models are large and involve different time scales, and it is time-consuming and data-demanding
to simulate the dynamic behaviors over long time intervals.
Based on the idea of time scale decomposition, the quasi
steady-state (QSS) [4] [6] seeks to reach a good compromise
between accuracy and efficiency. However, there are certain
limitations of the QSS model such as singularity problem.
When this happens, the Newton iterations diverge in practice
and the simulation cannot proceed. There are several papers
that addressed the singularity problem and tried to solve
it by a combination of detailed simulation and the QSS
approximation [7], Newton method with optimal multiplier [8],
and continuation method [9].
However, less attention has been paid to a severe situation
when the assessment based on the QSS model is not reliable.
Xiaozhe Wang is with the Department of Electrical and Computer Engineering, Cornell University, Ithaca, NY, 14853 USA e-mail: xw264@cornell.edu
Hsiao-Dong Chiang is with the Department of Electrical and Computer Engineering, Cornell University, Ithaca, NY 14853 USA email:hc63@cornell.edu

In this situation, the QSS model gives incorrect stability assessments in long-term stability analysis which means the QSS
model concludes the stability of the complete model, which is
in fact unstable. Due to the existence of these situations, the
QSS model may not consistently give conservative stability
analysis of the complete model. In other words, the QSS
model does not work under certain conditions, thus sufficient
conditions are needed under which the QSS model provides
correct stability assessment of the complete model.
This paper is organized as follows. Section II and Section
III briefly introduce the basic concept of complete dynamic
model and the QSS model with numerical examples. Section V
presents two counter examples in which the QSS model fails to
provide correct approximations of the complete model. Specifically, while the QSS model is stable, the complete model
suffers from voltage instabilities. Also, theoretical explanation
for this failure is presented. Conclusions and perspectives are
stated in Section VI.
II. C OMPLETE DYNAMIC M ODEL
The complete power system model for calculating system
dynamic response relative to a disturbance comprises a set
of first-order differential equations and a set of algebraic
equations [1]- [5]. The algebraic equations:
0 = g(zc , zd , x, y)

(1)

describing the electrical transmission system and the internal


static behaviors of passive devices. While the transient dynamics are captured by differential equations:
x = f (zc , zd , x, y)

(2)

which describe the internal dynamics of devices such as


synchronous generator and its associated excitation system,
interconnecting transmission network together with static load,
induction and synchronous motor loads, as well as other
devices such as HVDC converter and SVC. f and g are smooth
functions, and vectors x and y are the corresponding shortterm state variables and algebraic variables respectively. Both
continuous equations and discrete-time equations are needed
to represent long-term dynamics:
zc

zd (k + 1) =

hc (zc , zd , x, y)

(3)

hd (zc , zd (k), x, y)

(4)

where zc and zd are the continuous and discrete long-term state


variables respectively, and 1/ is the maximum time constant
among devices. These equations describe the dynamics of
exponential recovery load and thermostatically recovery load,

x
zc

= f (zc , zd , x, y)
= hc (zc , zd , x, y)

zd (k + 1) = hd (zc , zd (k), x, y)

(5)
(6)
(7)

Hence, the complete power system dynamic model involves


different time scales which makes the time domain simulation
over long time intervals very demanding. The QSS model
based on time-scale decomposition is proposed in [4] [6] [10]
and will be briefly stated in the following Section.
III. Q UASI S TEADY-S TATE M ODEL
The Quasi Steady-State (QSS) model is derived using the
idea of time-scale decomposition and aims to offer a good
compromise between the efficiency and accuracy [6]. In the
QSS model, the differential equations describing transient
dynamics are replaced by their equilibrium equations under
the assumption that transient dynamics are stable and settle
down infinitely fast in the long-term time scale.
Table I illustrates the concept of time-scale decomposition.
The transient model is obtained by assuming that slow variables zc and zd are constant parameters. While in the QSS
model, the transient dynamic equations (2) are replaced by
the corresponding equilibrium equations:
f (zc , zd , x, y) = 0

(8)

TABLE I

transient model
(approximation for transient stability)
short-term:0-30s
QSS model
(approximation for long-term stability)
long-term:30s-a few minutes

the trajectory comparison of algebraic variable V at Bus 3

2.9

longterm stability model


QSS model

2.85

1.016
complete model
QSS model
1.014
1.012

2.8
1.01

2.75
1.008

2.7

T HE MATHEMATICAL DESCRIPTION OF MODEL FOR POWER SYSTEM


complete model

model. This paper focus on the accuracy and reliability of the


QSS model instead of efficiency, thus the same time step as
that of the complete model will be used and the Jacobian of
the QSS model is updated at every time step as the complete
model.
A numerical example is presented below which shows the
trajectory comparison between the complete model and the
QSS model. The QSS model and the complete model finally
settle down to the same long-term stable equilibrium point
(SEP) in this case, and the QSS model provides a good
approximation of the complete model in long-term stability
analysis. However this is not always true which can be seen
from the counter examples presented in Section V.
The numerical study was performed using PSAT 2.1.6 [15]
on a modified model of IEEE 14-bus test system whose
one-line diagram is attached in Appendix A. There was a
fault at Bus 9 at 1s and the fault was cleared at 1.083s
by opening the breaker between Bus 10 and Bus 9. In the
complete model, the fast variables settled down by 30s after
the contingency, while the dynamics of load tap changer,
turbine governors and exponential load evolved in a longer
time period. The QSS model was used starting from 30s
when transient dynamics almost settled down. When the QSS
model was used, fast variables converged infinitely fast when
slow variables evolved. Finally, both the QSS model and the
complete model converged to the same long-term SEP. The
comparison of trajectories of the complete model and the QSS
model are shown in Fig. 1.

vr1 of Exc 2

turbine governor, LTC, OXL and armature current limiter, as


well as shunt capacitor/reactor switching all belongs to longterm dynamics. Note that shunt switching and LTC are typical
discrete components captured by Eqn (4).
Usually, transient (model) dynamics have much smaller time
constants compared with those of long-term dynamics, as a
result, zc and zd are also termed as slow state variables, and
x are termed as fast state variables. If we represent the above
equations in time scale where = t, and we denote as
d
d , then we have:

1.006

30

90

time(s)

150

200

1.004
0

30s(0.6)

60.6

120.6

170.6

zc = hc (zc , zd , x, y)
zd (k + 1) = hd (zc , zd (k), x, y)
x = f (zc , zd , x, y)
0 = g(zc , zd , x, y)
x = f (zc , zd , x, y)
0 = g(zc , zd , x, y)

Fig. 1. The trajectory comparisons of the complete model and the QSS
model for different variables. The trajectory of complete model followed that
of the QSS model until both of them converged to the same long-term SEP.

zc = hc (zc , zd , x, y)
zd (k + 1) = hd (zc , zd (k), x, y)
0 = f (zc , zd , x, y)
0 = g(zc , zd , x, y)

IV. N ONLINEAR F RAMEWORK : S TABILITY R EGION

Under certain conditions, the QSS model performs quite


well with similar accuracy as the detailed complete model,
while it takes much less time to simulate if a larger time step
or adaptive time steps are implemented. Also, compared with
the complete model, the Jacobian matrix of the QSS model
does not need to be updated at every time step, and it can be
updated only following discrete events such as LTC or OXL
activation unless slow convergence rate is observed [4]. As a
result, the QSS model is faster to simulate than the complete

Before presenting numerical examples, relevant definitions


are needed to give a theoretical explanation of the simulation results. If we are interested in the study region Uc =
Dzc Dzd Dx Dy , both models have the same set
of equilibrium points, that is E = {(zc , zd , x, y) U :
zd (k + 1) = zd (k), hc (zc , zd , x, y) = 0, f (zc , zd , x, y) =
0, g(zc, zd , x, y) = 0}. Assuming (zcls , zdls , xls , yls ) E is
an asymptotically long-term SEP of both the QSS model and
the complete model starting from (zc0 , zd0 , x0 , y0 ), and let
c (, zc , zd , x, y) be the trajectory of the complete model and
q (, zc , zd , x, y) be the trajectory of the QSS model starting
from the same initial point, then the stability region for the

complete model are defined as:


Ac (zcls , zdls , xls , yls ) := {(zc , zd , x, y) U : c (, zc ,
zd , x, y) (zcls , zdls , xls , yls ) as +}

(9)

For the QSS model, its dynamics are constrained to the set:
:= {(zc , zd , x, y) U : f (zc , zd , x, y) = 0, g(zc , zd , x, y) =
0} which is termed as the constraint manifold. Note that the
constraint manifold may not be smooth due to the discrete
behavior of zd . Then the stability region of (zcls , zdls , xls , yls )
for the QSS model are defined as:

c (, zc , zd , x, y) of the complete model will not converge


to the long-term SEP (zcls , zdls , xls , yls ) which the trajectory
q (, zc , zd , x, y) of the QSS model converges to. Hence,
the QSS model is not an appropriate approximation for the
complete model and gives incorrect stability assessments in
this case.
family of transient models

Aq (zcls , zdls , xls , yls ) := {(zc , zd , x, y) : q (, zc ,


zd , x, y) (zcls , zdls , xls , yls ) as +}
(10)
Similarly, for the transient model with fixing slow variables
zc and zd (k):
x = f (zc , zd (k), x, y)
0 = g(zc , zd (k), x, y)

(11)

the equilibrium points are termed as transient SEPs. The


stability region of transient SEP (zc , zd (k), xts , yts ) is defined
as:
At (zc , zd (k), xts , yts ) := {(x, y) Dx Dy , zc = zc ,
zd = zd (k) : t (t, zc , zd (k), x, y) (zc , zd (k),
xts , yts ) as t +}
(12)
where t (t, zc , zd , x, y) is the trajectory of the transient model
(11). A comprehensive theory of stability regions can be found
in [11] [12] [13] [14].
Generally, the SEPs of each transient model are isolated
and the trajectory q (, zc , zd , x, y) of the QSS model does
not meet the singular surface and is constrained on s all the
time where s is defined as:
s = {(zc , zd , x, y) : all eigenvalues of(

f
f

x
y

g 1 g
) satisfy Re() < 0, g/y is nonsingular}
y x
(13)
Note that each point of s is a SEP of the transient model
defined in Eqn (11) for fixed zc and zd (k). Thus generally
the trajectory q (, zc , zd , x, y) of QSS model moved along s
on which each point is a SEP of the corresponding transient
model. Given enough simulation time which is usually to be
several minutes, both the QSS model and the complete model
converge to the same long-term SEP.
However, if when zd firstly change from zd (k 1) to
zd (k), and the initial point (zc , zd (k), x0 , y0 ) on the trajectory c (, zc , zd , x, y) lies outside the stability region
At (zc , zd (k), xts , yts ) of the transient model:
x =

f (zc , zd (k), x, y)

0 =

g(zc , zd (k), x, y)

(14)

then c (, zc , zd , x, y) will move away from the slow manifold s as shown in Fig. 2. As a result, the trajectory

the long term SEP

asymptotically SEPs
of transient models

Fig. 2. When zd firstly change to zd (k), the initial point of the complete
model get outside of the stability region of the transient model and the
trajectory of the complete model moves far way from the QSS model from
then on.

V. C OUNTER E XAMPLES
The QSS model has some limitations in dealing with severe
disturbances. As stated in [4], the QSS model cannot reproduce
the instabilities where the slow variables trigger instability of
fast variables. This means the QSS model can not capture the
insecure cases when the fast variables are excited by the slow
variables, thus result in voltage instabilities. In addition, the
QSS model may converge to another stable equilibrium point
different from the one the complete model converges to. Under
these two situations, the QSS model does not capture the
dynamic behavior of the complete model and give inaccurate
approximations of the complete model. In brief, the QSS
model can lead to incorrect stability assessment.
A. Numerical Example I
This system was set up based on the modified IEEE-14 bus
system in Section III. Apart from the two turbine governors
at Bus 1 and Bus 2 , there were three exponential recovery
loads at Bus 9, Bus 10 and Bus 14 respectively, and five over
excitation limiters were added for each exciter which started
to work after a fixed delay 10s. Besides there were three load
tap changers which are discrete models [4]:

mk + m if v > v0 + d and mk < mmax


mk m if v < v0 + d and mk > mmin
mk+1 =

mk
otherwise
(15)
where m denotes the lap changer ratio. The one-line diagram
of the modified system is also attached in Appendix A.
There were two faults at Bus 9 and Bus 6 that happened
simultaneously at 0.02s, and the faults were cleared by opening
the breakers between Bus 7 and Bus 9, between Bus 6 and
Bus 11 at 0.1s, and the one between Bus 6 and Bus 13 at

1s. The complete model was employed for the first 30s while
the QSS model was employed afterwards. The comparison of
trajectories of different variables in the complete model and
the QSS model is showed in Fig. 3.
In this case, the QSS model failed to give a correct approximation of the complete model. The time domain simulation
of the complete model stopped and stated that there was singularity likely in the system around 101.2155s (71.8155 ),
while the QSS model did not encounter such problems and
continued to converge to the long-term SEP. From Fig. 3, it
can be seen that in the complete model, fast dynamics x were
excited when slow variables evolved. The violent variation
of fast variables x due to slow variables finally resulted in
voltage instability of the complete model such that it did not
converge to the same asymptotically SEP as the QSS model.
However, if we only look at the QSS model, the dynamics
of fast variables x due to slow variables are not noticeable
since x and y converged to the transient SEPs immediately.
Therefore, if the state of long-term SEP is acceptable, the postfault system will be misclassified as stable. In this case, the
assumption behind the QSS model that transient dynamics are
stable in long-term time scale is violated.

sponding transient model starting from (zc , zd (2), x0 , y0 ):


x =

f (zc , zd (2), x, y)

0 =

g(zc , zd (2), x, y)

is plotted in Fig. (5). It can be seen that both the fast variable
and the algebraic variable converged to the SEP of the transient
model (17). In other words, the initial point (zc , zd (2), x0 , y0 )
of the complete model (16) is inside the stability region
At (zc , zd (2), xts , yts ) of the transient model (17).
However when zd changed from zd (2) to zd (3) at 40s, the
complete model was no longer stable which can be seen from
Fig. (6). The fast variables were excited by the evolution of
slow variables zd and zc . The trajectories of fast variables in
the corresponding transient model are plotted in Fig (7), and
the initial point (zc , zd (3), x0 , y0 ) of the complete model (16)
(substitute zd (2) by zd (3)) was outside of the stability region
At (zc , zd (3), xts , yts ) of the transient model (17) (substitute
zd (2) by zd (3)) As a result, the QSS model gives incorrect
approximations of the complete model from then on.
the trajectory comparison of transient variable v of Exc 2
r1

3.75
complete model
QSS model

3.7
the trajectory comparison of fast variable of Syn 2
0.3
complete model
QSS model

the trajectory comparison of transient variable v of Exc 1


r1

2.5
complete model
QSS model
2

0.35

1.055
1.05
1.045

3.4
1

0.45

0.5
0 30s(0.6)

3.6
3.55
3.5

0.4

3.35
0

20

40

60

80

100

1.04
0

20

40

60

80

100

0.5

60.6

120.6

170.6

the trajectory comparison of slow variable v

oxl

of Oxl 2

0.25
0.2

0
0 30s(0.6)

60.6

120.6

170.6

the trajectory comparison of algebraic variable V at Bus 1


1.15
complete model
QSS model
1.1

0.15

1.05

0.1

Fig. 4. The trajectories comparisons of the complete model and the QSS
model for different variables when load tap changers changed at 30s. Both
the complete model and the QSS model converged to the same SEP.

3.68

1.065
the trajectory of transient variable vr1 of Exc 2

3.67

0.95

0.05
complete model
QSS model
60.6

120.6

170.6

1.066
1.0665

3.65

1.067
3.64

0.9
0 30s(0.6)

1.0675

60.6

120.6

170.6

3.63
1.068
3.62

1.0685

3.61

Fig. 3. The trajectory comparisons of the complete model and the QSS model
for different variables. The assumption of QSS model that the fast variables
are stable is not satisfied such that it gives incorrect approximations.

The failure of the QSS model can be further explained by


checking the trajectory of the transient model. When zd firstly
changed from zd (1) to zd (2) at 30s, denote the initial point
on the trajectory of the complete model when this change
happened as (zc , zd (2), x0 , y0 ), then the complete model fixed
at zd (2) starting from (zc , zd (2), x0 , y0 ):
zc

x =
0 =

the trajectory of algebraic variable at Bus 9


at Bus 9 at the SEP of the transient model

1.0655

vr1 of Exc 2 at the SEP of the transient model

3.66

0
0 30s(0.6)

the trajectory comparison of algebraic variable V at Bus 2


1.065
complete model
QSS model
1.06

3.65

3.45

1.5

(17)

hc (zc , zd (2), x, y)

(16)

f (zc , zd (2), x, y)
g(zc , zd (2), x, y)

was stable as shown in Fig. (4) in which both the complete


model and the QSS model converged to the same long-term
SEP. Moreover, the trajectories of two variables in the corre-

1.069

3.6
3.59

1.0695
0

10

1.07

Fig. 5.
The trajectories of the transient model when load tap changers
changed at 30s which indicated that (zc , zd (2), x0 , y0 ) was inside the
stability region of the transient model.

B. Numerical Example II
Another numerical example performed on a modified IEEE
145-bus system is presented below. Due to limited pages, only
simulation results are shown in Fig. 8. We can see that the
voltage at Bus 90 was collapsed around 235s in the complete
model, however, the voltage at Bus 90 settled down to the
value around 0.9344 p.u in the QSS model. Also, the QSS
model did not provide correct approximations for transient
variables.

10

the trajectory comparison of transient variable vr1 of Exc 2

3.8

x 10 the trajectory comparison of longterm variable v of Oxl 2


complete model
QSS model

Bus 13

Bus 13

3.7

Bus 14

Bus 14

Bus 10

Bus 10

Bus 12

Bus 12

3.6

Bus 09

Bus 09

Bus 11

Bus 11

3.5

Bus 07

Bus 07

Bus 06

3.4

complete model
QSS model

3.3
0

20

40

60

80

Bus 06

Bus 04

20

40

60

80

100

Bus 05

Bus 01

Bus 01

Bus 02

Fig. 6. The trajectories comparisons of the complete model and the QSS
model for different variables when load tap changers changed at 40s. The
complete model was unstable while the QSS model converged to a SEP.

3.66

3.78

3.655

3.77

Bus 02

Bus 03

Bus 03

Fig. 9. One-line diagram of the example in Section III; One-line diagram


of the example in Section V-A

the trajectory of transient variable vr1 of Exc 2

3.65

Bus 04

Bus 05

1
0

100

Bus 08

vr1 of Exc 2 at the SEP of the transient model

3.76

3.645

3.75

3.64

ACKNOWLEDGMENT

3.74
3.635
3.73
3.63
3.72

3.625

3.71

3.62
the trajectory of transient variable vf of Exc 2
3.615

3.7

v of Exc 2 at the SEP of the transient model


f

3.61

3.69

10

Fig. 7.
The trajectories of the transient model when load tap changers
changed at 40s which indicated that (zc , zd (3), x0 , y0 ) was outside of the
stability region of the transient model.
the trajectory comparison of algebraic variable V at Bus 90
1

the trajectory comparison of transient variable vr1 of Exc 2


2.65

0.8

2.6

0.6

2.55

0.4

2.5

0.2

2.45

complete model
QSS model
0
0 40s(0.8)
80.8

160.8

260.8

complete model
QSS model

0 40s(0.8)

80.8

160.8

260.8

Fig. 8. The trajectory comparisons of the complete model and the QSS
model. The QSS model converged to a long-term SEP while the complete
model suffered from voltage collapse.

VI. C ONCLUSION
The QSS model was derived based on time-scale decomposition and it offers a good compromise between accuracy
and efficiency. In this paper, two counter examples in which
the QSS model provides inaccurate stability assessments are
presented, and the reasons for the inability of the QSS model
to approximate the complete model are explained from the
stability regions of the transient models of the complete model.
These counter examples suggest that there is a necessity to
provide a theoretical foundation for the QSS model. Moreover,
an improved QSS model may be needed in order to give
consistently accurate approximation of the complete model.

A PPENDIX A
T HE O NE -L INE D IAGRAM OF N UMERICAL E XAMPLES
The one-line diagram of the numerical examples are shown
in Fig. 9.

10

The authors would like to thank Dr. Luis F. C. Alberto for


helpful discussions. And this work was partially supported by
the CERT through the National Energy Technology Laboratory
Cooperative Agreement No. DE-FC26-09NT43321.
R EFERENCES
[1] H. D. Chiang, Direct Methods for Stability Analysis of Electric Power
Systems-Theoretical Foundation, BCU Methodologies, and Applications.
New Jersey: John Wiley & Sons, Inc, 2011.
[2] P. Kundur, Power System Stability and Control. New York: McGraw-Hill,
Inc. 1994.
[3] P. Kundur, J. Paserba, V. Ajjarapu, Definition and Classification of Power
System Stability. IEEE Transactions on Power Systems, Vol. 19, No. 2,
pp. 1387-1401, May 2004.
[4] T. V. Cutsem, Voltage Stability of Electric Power Systems.
Boston/London/Dordrecht: Kluwer Academic Publishers, 1998.
[5] P. W. Sauer, M.A. Pai, Power System Dynamics and Stability. PrenticeHall, New Jersey, U.S.A. 1998.
[6] T. V. Cutsem, Y. Jacquemart, J. N. Marquet, A Comprehensive Analysis
of Mid-term Voltage Stability. IEEE Transactions on Power Systems, Vol.
10, No. 3, pp. 1173-1182, August 1995.
[7] T. V. Cutsem, M. E. Grenier, D. Lefebvre, Combined Detailed and
Quasi Steady-State Time Simulations for Large-disturbance Analysis.
International Journal of Electrical Power and Energy Systems, Vol. 28,
Issue 9, pp. 634-642, November 2006.
[8] P. Rousseaux, T. V. Cutsem Quasi Steady-State Simulation Diagnosis
Using Newton Method with Optimal Multiplier. Power Engineering
Society General Meeting, 2006.
[9] Q. Wang, H. Song, V. Ajjarapu Continuation-Based Quasi-Steady-State
Analysis. IEEE Transactions on Power Systems, Vol. 21, No. 1, February
2006.
[10] L. Loud, P. Rousseaux, D. Lefebvre, T. V. Cutsem A Time-Scale
Decomposition-Based Simulation Tool for Voltage Stability Analysis. In:
Proceedings of the IEEE Power Tech Conference, Vol. 2, Porto, Portugal,
2001.
[11] H. D. Chiang, M. W. Hirsch, and F. F. Wu, Stability regions of nonlinear autonomous dynamical systems. IEEE Transactions on Automatic
Control, vol. 33, no. 1, pp. 1627, Jan. 1988.
[12] H. D. Chiang, J. S. Thorp, Stability regions of nonlinear dynamical
systems: A constructive methodology. IEEE Transactions on Automatic
Control, vol. 34, no. 12, pp. 12291241, December 1989.
[13] J. Zaborszky, G. Huang, B. Zheng, and T. C. Leung, On the phase
portrait of a class of large nonlinear dynamic systems such as the power
system. IEEE Trans. on Automatic Control , vol. 33, no. 1, pp. 415,
January 1988.
[14] Luis F. C. Alberto, H. D. Chiang, Theoretical Foundation of CUEP
Method for Two-Time Scale Power System Models. Power and Energy
Society General Meeting, 2009.
[15] F. Milano Power System Analysis Toolbox Documentation for PSAT
version 2.1.5, November 1, 2009.

Bus 08